SparkFun Forums

[geetar_king]

I am trying to design something that tracks sea level height at a buoy location.

I would like to have an enclosed box attach to a buoy and have the internals allow the recording of the vertical displacement of the buoy. If the vertical coordinate is Z then I do not, for example, care to record or track X or Y displacement although the buoy may make slight movements in these X, Y directions. I only need to track Up/Down movement of buoy to give me an indication of water level.

Can someone advise what the best way to do this would be? Can I accomplish this with an IMU? How many degrees of freedom would I need. Or a force sensor something?

[waltr]

It might be possible with an IMU and maybe even possible with just an accelerometer.
However, you will need all three axes since I'm sure the buoy will not stay level.

The other problem if you're trying to track very slow changes, like tides, is accumulated error in the integration.

How about Sonar to measure the distance to the sea bottom.

[geetar_king]

Yeah I was wondering if the low frequency / slow movement would be an issue. Sonar is an option, but I would like to try without so that the device could be applied also to ground level change if needed. Ie seasonal ground heave, which would have the device sit directly on the ground.

Is there a way with IMU or accelerometer to get away from the error accumulation? Multiple accelerometers and averaging?

[Mee_n_Mac]

How tightly is the buoy tied to the bottom ? If the water level rises, does the buoy have enough slack to rise with it or does the water rise around the buoy ?

[geetar_king]

The buoy would rise with water level

[geetar_king]

Which is the best accelerometer for low frequency / slow acceleration? I'm guessing I should be looking for lowest noise specification?

[waltr]

Start by searching use of IMUs and read up on to techniques used and issues with accumulated errors.
The basic issue is measurement from an accelerometer are taken periodically. One then uses the acceleration of that measurement and 'assumes' it was constant since the last measurement. This is 'integrated' (calculus operation) to obtain a displacement since the last measurement and added (+ or -) to the last position.
The problem is the assumption that the acceleration is constant between measurements. If it is not (very likely) then there is a position error on each measurement that accumulates.

[Mee_n_Mac]

I don't see how any IMU you can afford is going to work. You mentioned seasonal changes, implying months of measurement time ? RTK GPS would be 1 method, sonar or a dropping line (think old time mariners) would be 2 more.

[lyndon]

The simplest thing I can think of that is even remotely likely to work is differential GPS.

[stevech]

I'd use a simple MEMs device with a scale-range of 1G or less.

[Mee_n_Mac]

Another thought is to put a pressure sensor on the bottom under the buoy. Then, by wire or acoustics, relay bottom pressure to a recording box on the buoy. Depth can be known from (bottom pressure-atmospheric pressure)/density of seawater. How deep is the sea going to be by the buoy(s) ?

[RonBeyer]

http://www.codefurnace.com/Man_user_mru_r11.pdf

That is a detailed document showing the design of an MRU (motion reference unit), specifically the measurement of heave. Its a sensor specifically designed to measure vertical displacement with high resolution. It also costs $30,000...

[stevech]

there are Tsunami warning boys that measure slow steady rises in sea level.

Maybe a cheap version is a mass on a line on the sea floor. Measure line (played out) length. Calibrate-out tides.

[fll-freak]

The mass on the end of a line might work for a week before marine life would build up. I have seen anchor chain so built up with grown as to render it stiff.

GPS is out. Vertical height is notoriously bad.

Baro pressure is out as day to day changes would be orders of magnitude larger than the signal.

An IMU would be near hopeless for long duration measurements. Any IMU has a drift rate due to time and temperature you would have to zero out. A MEMS based IMU may keep its accuracy for a few minutes. A ring laser based IMU perhaps for an hour. But no IMU that I am aware of could measure to a few feet over days of operation. So you would need a way to reset the zero every hour in the best case.

Sea bed floor pressure has a chance. Problem here is cabling to the surface. You could run it on batteries and send the data up accousticaly. When the batteries get low, have it cut it's own anchor and float to the surface to be replaced. Copper cabling has all the problems of marine grown and servicing problems.

Sonar/laser might work IF the sea bed floor was very flat. If not, as the buoy sifted horizontal position, it would report different heights.

How far out to sea is this buoy?
Could you use and automated transit system to survey the altitude from land?
What is your budget?
What is the time span for measurement?
What is the needed accuracy?

The devil is in the details.

[lyndon]

I said differential GPS. DGPS is used to position the blades of earthmoving equipment with an accuracy of centimeters when used for final grading. Here is a table of accuracies of some off the shelf equipment. Sub-centimeter accuracies are possible.